04_Assignment.Rmd

---
title: "Assignment"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Introduction
For this assignment, you will use The Cancer Genome Atlas (TCGA) glioma RNA-sequencing data. This dataset is a matrix of RSEM normalized and log2-transformed gene expression values. This dataset consists of 1,129 samples and was retrived from Broad GDAC Firehose (https://gdac.broadinstitute.org/).

Here, you're tasked to calculate the coefficient of variance (CV2) for each gene. The CV2 is calculated by dividing variance by the square of mean. CV2 is often used as a measurement of variability, and typically genes with moderate-to-high CV2 are brought forward for downstream analysis such as dimension reduction and clustering.

Specifically, you need to benchmark the different approaches taught in _Efficient R Codes_ for calculating CV2, namely:

- `for` loops that grow vectors
- `for` loops that don't grow vectors
- Utilizing apply family
- Parallel computing
- Vectorize code

## Getting started
```{r}
# Load packages
library(data.table)
library(parallel)
library(microbenchmark)

# Read file
df <- fread("Datasets/GBMLGG.uncv2.mRNAseq_RSEM_normalized_log2.txt", sep="\t", header=TRUE, stringsAsFactors=FALSE)

# Check dimensions
dim(df)

# Sneak peek
df[100:105,1:5]
```